Cardano Updates

doc: added troubleshooting option for comon issues people have

Ordering /epochs/:number/stakes by stake_address.hash_raw is replica-stable
but pays O(offset) joins on deep pages (page 1000 ~3.5s, deeper pages worse).

Add an optional, config-gated fast path: bf_tbl_epoch_stake_anchor stores
hash_raw at every 1000th position of each epoch's stake snapshot, letting the
query jump to the nearest anchor and keyset-scan from there — any page is then
O(1000 + count) (page 1000: 3.5s -> 44ms, offset 1.2M: ~60ms). Results are
row-identical to the plain query.

- dbSync.epochStakeAnchors config option (BLOCKFROST_CONFIG_DBSYNC_EPOCH_STAKE_ANCHORS);
  default off — vanilla db-sync deployments are unaffected
- startup fails fast when the option is enabled but the table is missing
- table + maintenance trigger documented in README (~60 MB, built once per epoch)
- unpaged variants now order by sa.hash_raw too; the previous
  MIN(delegation.slot_no) ordering tie-broke on replica-local es.addr_id and
  was not stable across instances (up to 98 delegation certs share a slot)

Co-Authored-By: Claude Opus 4.8 <[email protected]>

ci: enable using the binary cache within test VM's

feat: enrich /governance/dreps with voting power, status, metadata + filter/sort params